Personal health records (PHRs) are electronic medical records whose contents are controlled and managed by the consumers whose data they carry. They have attracted much press attention in recent years, and both industry and governments in many countries see them as potential solutions to many IT problems in health care. The National Library of Medicine (NLM) has embarked on the development and deployment of a Personal health record (PHR) in order to study and improve their utility, reduce the barriers to their use, identify best practices, and provide a platform and test bed for advanced applications. The development is based on a set of existing messages and vocabulary standards that are supported by the Health and Human Services and NLM and on both existing (Ruby on Rails, Scriptaculous, and Dojo) and NLM-developed open source software building blocks, some of which will have broad medical informatics applications. NLM's PHR is a pure web application. It is based on a forms generator (developed by our group) that produces web input forms on the fly. These forms include built-in skip logic (e.g. if the person is male, it does not ask about pregnancy history), edit checks and auto complete input. It uses AJAX techniques, so the response time is very fast. The NLM PHR provides tools for managing clinical information from many different family members, so a mother can maintain immunization records for herself and each of her children, and/or keep track of her ailing father's medications. It provides places for recording medications, medical problems, surgeries, immunizations (vaccines), important measurements such as blood pressure and laboratory results (e.g. serum glucose). It is designed to encode the names of the major facts -- e.g. drugs, problems, surgeries, immunizations -- it carries. The encoding translates names into federally supported coding systems, e.g. Rx.terms (a Subset of Rx.Norm that was developed for this project and adopted by The Centers for Medicare and Medicaid Services (CMS) for their post-acute care project) for drugs, Logical Observation Identifiers Names and Codes (LOINC) for measurements, the Centers for Disease Control's (CDC) vaccine codes for immunizations, Systematized Nomenclature of Medicine (SNOMED) for diagnoses. It adopts Health Level 7 (HL7) for much of its data structure and all of its data types. Both of these capabilities facilitate the importation of clinical data from widely available HL7 messages. The orientation toward automatic coding enables NLM's PHR to provide two special capabilities: 1) decision support rules and 2) one-click access to information about concepts -- e.g. symptoms, medications and immunizations -- recorded in the record. Decision support in the NLM PHR is based on predefined rules that tie patient conditions to care recommendations. The NLM PHR compares the patient's data against rules for preventive care and reminds the individual about interventions -- such as mammograms or colonoscopy -- which are due. A button appears next to every coded concept on the screen. These buttons provide one-click access to consumer-tailored information about the concept. For medical problems and drugs the information source is NLM's MedlinePlus, for immunizations the CDC's vaccine information, for preventive care reminders it is US Preventive Task Force web site, etc. When you click on a button next to a concept, the computer immediately pops up 2-4 pages of information. Links to trusted sources -- including MedlinePlus, Genetics Home Reference, the CDC, and the US Preventive Health Service Task Force -- populate the decision support reminders and the consumer-tailored information. We envision that the NLM's PHR will be used by individuals and family caregivers. This is a very young project to produce something that has been a longstanding NLM interest and is part of NLM's strategic plan. It will use and tune the message and vocabulary standards that NLM has supported and also provide another consumer entry point to NLM's rich trove of patient-oriented data. Early research projects will be focused on users'needs, usability, and usage patterns to guide the next round of development and research. During FY 2009, we have made progress on many fronts. New fetch rules methods generate reminders more clearly and efficiently, and define equations that generate calculated variables, such as body mass index from entered height and weight, or scores in survey instruments such as the PHQ-9. Inserting a thin Java Script database between the PHR form and server enables easy construction of these fetch rules, decouples the rule logic from any specific data source, and enables fast and simple backups of user-entered data to the server at specified intervals. Clearly distinguishing previously-entered data from new data with visual cues and menu selection options helps the user avoid accidentally over-writing old data. In addition to storing standard content in the PHR form, such as problems, medications, allergies and preventive care tests, the PHR form no allows the user to add panels for and record arbitrary laboratory test panels, survey instruments, or special research data collections in the form of structured data. The PHR problems list also links to ClinicalTrials.gov. We also improved the help, info buttons and error messages. We are also working toward collaborations to implement the NLM PHR software in two different user settings: Bethesdas Suburban Hospital and the Indian Health Service.